Long-Term Video Question Answering via Multimodal Hierarchical Memory Attentive Networks | IEEE Journals & Magazine | IEEE Xplore